Homebrew 3 Channel Pure SIN Generator ( 0 ... 60MHz )

[otto]

Hello all,

something very important for me. As I cant use PMs I have to do it in this way:

@tao

years ago you posted to discharge caps into the TPU or something like this. We were told that this is SMs secret.

Thanks for this.

Otto

[Yucca]

KISS

I agree. That's why I'm using off the shelf modules and plugging them together like lego, what could be simpler! Why reinvent the wheel when a group of radio HAMs have evolved a really nice oscillator module already over several years, you can get each module as a kit for only $30 and then you get a free sample chip. Or you can buy the modules preassembled.

@Peter, the EEPROM has 100k write cycles guaranteed, so less hardy than pic. I suppose If I just dump state to beginning blocks then it should last a lifetime if it only writes on powerdown. My caps are a whopping 1F and it has a few seconds on powerdown because the micro has a 5V reg so the 10V supply can afford to drop to 6V.

[Yucca]

@Otto,

I am not an expert on the TPU, definitely not! You may be right, no it may not be needed for working TPU but it will help me learn more about radio like systems.

I plan to look for weird effects in lots of systems not just the TPU. I think 3 channels is overkill, I was going to do just 2, but its always good to have a spare 

So now you know I am only building a 3 channel radio quality oscillator. I am spending 3 months hobby budget on this thing but I hope to have it for many years.

Yucca.

[giantkiller]

Yes! This the way to get the skin effect before the conductor conducts!
Now what if you have a bias field around a conductor and then shock that? Canon balls, my friends, canon balls. Directional travelling shock waves.

Grumpy posted Ehrenhaft.
http://www.rexresearch.com/ehrenhaf/ehrenhaf.htm
I noticed this diagram and related it to the TPU. I can't be that far off.

Hello all,

something very important for me. As I cant use PMs I have to do it in this way:

@tao

years ago you posted to discharge caps into the TPU or something like this. We were told that this is SMs secret.

Thanks for this.

Otto

[Yucca]

@all (including all miserable old farts): I've just increased the GUI resolution to 100mHz. The oscillator modules have an incredible 24mHz resolution but one tenth Hz seems good enough. I had to reduce the digit widths by 2 pixels but it is still easily readable up to 1m away.

I've been in correspondence with the module designers, they assure me I should have no problems by sharing a common clock using same length runs of RG174 coax. I may also have to have measured lines to the load signal inputs. But I will try on the same bus to start with as only 1cm length between modules. If I cant assert within refclock*6 on all lines then I'll also have to use same length coax to route that signal also.

Also just picked up a bucketload of SIL pin plugs and sockets from ebay china for peanuts, shame about the 3 week wait, but no problem because I have a lot of software to write.

I had a worry today because my LCD kept crashing after many fast writes to it, I spent ages looking for a code error and now I realise that when powered by USB cable the power getting to the screen is marginal. When I power 10V through the regulator everything is rock solid.

Any ideas about menu fuctions would be welcome before I start to code, at the moment I'm thinking:

[1]
linear sweep (1,2 or 3D) (maximum ADC signal shown at end)

[2]
log10 sweep (1,2 or 3D) (maximum ADC signal shown at end)

[3]
1D graphical spectrum analyser (realtime ADC response graph plotted)

[3]
2D graphical spectrum analyser (realtime ADC response graph plotted)